| Objective:Liver cirrhosis is a disease that seriously threatens the health of residents in China.And its prevalence is high and it is difficult to reverse the disease after illness.When the disease progresses to the decompensation period,it can cause a variety of complications,the mortality rate rises significantly,eventually requires liver transplantation,and the prognosis is extremely poor.Therefore,a reliable scoring system is needed to assess the liver damage and prognosis of patients with cirrhosis,so that providing a basis for prognosis and clinical treatment choice.Child-Pugh classification is a widely used scoring system for evaluating liver function.However,this scoring system has the disadvantages of a narrow grading range,limited ability to discriminate,and the same weight for all variables.In order to make up for the shortcomings of the traditional Child-Pugh scoring system,the latent structural model was applied to the Child-Pugh scoring system of liver function.It can help us get a better liver function grading system,so as to improve the prediction effect of cirrhosis classification,identify the patients with high-risk cirrhosis of liver damage,through early intervention and the preventive treatment measures,reduce the fatality rate of liver cirrhosis.Methods:By retrospectively collecting medical records of 1561 cirrhotic patients admitted to the Second Hospital of Shanxi Medical University from January 2010 to April 2017,latent trait analysis and the ordered latent class analysis were carried out for the score of5 ordered indicators(namely hepatic encephalopathy,ascites,albumin,prothrombin time and total bilirubin)in the Child-Pugh scoring system.Through latent trait analysis,wegot a comprehensive variable with weight-the principal component scores.The t-test was used to compare the scores of the principal components of the two groups whether there were adverse outcomes or not,and the correlation analysis between the principal component scores and the surface scores was carried out.And then,The predictive value of traditional CTP scores and principal component scores for adverse outcome of cirrhosis was compared using the area under the ROC curve.Latent class analysis was used to identify patients with different liver damage and chi-square test was used to verify that latent class analysis can identify patients with severe liver damage.Then,evaluating the consistency between the latent class analysis and the traditional CTP classification.The predictive value of traditional CTP classification and latent class analysis for adverse outcome of cirrhosis was compared using the area under the ROC curve.Finally,an optimal grade scoring system for liver cirrhosis is obtained and verified.Results:1.Latent trait analysis showed that the principal component scores of all response modes were between(8.167~24.501).As the score of different response mode increased,the principal component score also increased,indicating that the higher the severity of liver damage.The results of the principal component score showed that albumin had a great influence on the severity of liver damage,prothrombin time was the second,and total bilirubin was smaller.The principal component score was positively correlated with the surface score and had a high degree of correlation.There were significant differences in the scores of the principal components between the occurrence and non-occurrence of adverse outcomes of cirrhosis(eg: death,hepatorenal syndrome,electrolyte disturbance,spontaneous peritonitis,upper gastrointestinal bleeding).The area under the ROC curve predicting the adverse outcome of liver cirrhosis was calculated by CTP scores and principal component scores.And the result showed that the two methods of predicting adverse outcome of cirrhosis makes sense(P<0.05).There was no significant difference in the predictive ability of CTP score and principal component scores for death,deathand critical illness,spontaneous peritonitis,electrolyte disturbance and hepatorenal syndrome,indicating that the CTP scores was not inferior to the principal component score,and the CTP scores was acceptable.2.Latent class analysis stratified patients with cirrhosis into three latent clusters:346(22.2%)with LCA-A(low severity of liver damage),596(38.2%)with LCA-B(moderate severity of liver damage),and 619(39.7%)with LCA-C(high severity of liver damage).The number of patients with high severity of liver damage was the most in this study.The chi square test between latent class analysis and the occurrence of adverse outcomes were all P <0.05,indicating a statistically significant difference in the incidence of adverse outcomes among patients with different latent clusters of cirrhosis,which further confirmed that the occurrence of adverse outcomes was associated with the latent clusters and the incidence of adverse outcomes was high in patients with severe liver damage.The consistency between the latent clusters and the traditional CTP classification was 0.826 and the Kappa value was 0.726.The area under the ROC curve predicting the adverse outcome of liver cirrhosis was calculated by CTP classification and latent class analysis.And the result showed that the two methods of predicting adverse outcome of cirrhosis makes sense(P<0.05).The predictive ability of CTP classification and latent class analysis on death,death and morbidity,hepatorenal syndrome,upper gastrointestinal bleeding,combined with five adverse outcomes were all latent class analysis>CTP classification,Which showed that CTP classification had a low predictive value and needed improvement.3.The predictive ability of CTP scores,CTP classification and latent class analysis on death were 0.838,0.714,0.754,Which showed that the cut-off point of CTP classification is unreasonable.Therefore,we will form a new grading standard by combining latent class analysis and CTP scores.We named it LCA-CTP classification and we formed a new threshold of 5~7 divided into mild,8~10 divided into moderate,11~15 divided into severe.P value of LCA-CTP classification was less than 0.05,indicating that it can predict adverse outcome of cirrhosis.The area under ROC curve forvarious adverse outcomes showed that LCA-CTP classification was superior to CTP classification,and the comparisons of the incidence of adverse outcomes with different LCA-CTP classification further confirmed it.Conclusion:This study,for the first time,combined the latent structural model with the Child-Pugh scoring system to improve the traditional CTP classification and proposed a new scoring system for liver cirrhosis that was superior to the traditional Child-Pugh classification,namely LCA-CTP classification.Finally we formed a new threshold of5~7 divided into mild,8~10 divided into moderate,11~15 divided into severe. |
PDF Full Text Request